The efficacy of gene therapy against brain tumors will depend upon delivery of viral vector throughout tumor and specific cytotoxicity toward both infected and non-infected tumor cells. Previously this project has focused on delivery, assessing interstitial infusion and transvascular delivery of particles by osmotic opening of the blood-brain barrier (BBB). While delivery remains our major focus, the current proposal will also move to a broader examination of virus and particle uptake and efflux. In addition to recombinant adenovirus vectors, as a model for virus particles we will use viral-sized iron oxide nanoparticles, Combidex and Code7228, because they allow direct comparison of magnetic resonance (MR) imaging with histology and ultrastructure. In Aim 1 we will assess influx and uptake of virus and iron particles into rat brain and intracerebral tumor, and investigate the effect of tumor size, permeability and prior irradiation. Since radiation may increase virus distribution and/or transgene expression, we will test a novel approach combining tumor-specific radioimmunotherapy with virus delivery. This aim will also evaluate iron particle influx in a rat stroke model, and characterize phagocytic and/or astrocytic reactive cells responsible for iron particle trapping. Aim 2 will evaluate virus and iron particle efflux from the brain, using mR and histology to delineate efflux pathways. In Aim 3, we will investigate the potential for gene therapy with Herstatin, a secreted protein which inhibits the epidermal growth factor receptor (EGFR). We will compare intratumor vs. transvascular delivery of protein and a Hestatin adenovirus construct, both in the LX-1 lung cancer metastasis model, as well as in glioma and breast metastasis models. We hypothesize that because Herstatin is secreted and has high activity against EGFR overexpressing cells, it will provide bystander efficacy even when a small proportion of tumor is infected. Finally, Aim 4 will be an expanded clinical protocol of rion particle imaging and localization, to assess BBB and blood-tumor barrier permeability and Combidex uptake in adult and pediatric brain tumors, and in CNS inflammatory lesions. We will also evaluate tumor vascularity using MR angiography with Code7228, a new formulation which allows bolus administration and dynamic MRA superior to Gd. Our hypothesis is that tumor imaging with the iron oxide particle agents requires both a leaky BBB and trapping by uptake into reactive cells. We anticipate that these studies will not only be useful in designing clinical trials of brain tumor gene therapy, but also in providing a new means to image CNS tumors, neurological lesions and even gene therapy approaches which evoke a cellular reaction.